Fix pre-SDK Cache_Read_Enable for PUYA flash (esp8266#8658)

* Enable SPI_CS_SETUP for early ICACHE use

The SPI_CS_SETUP parameter has been observed set by RTOS SDK and NONOS SDK
as part of flash init/configuration. It may be necessary for some flash
chips to perform correctly with ICACHE hardware access. Turning on and
leaving it on should be okay.

* Cleanup comment

* Change umm_init to default to IRAM

Some flash chips (PUYA) have some unknown requirements for running with
early `Cache_Read_Enable`. They work fine after the SDK is started.
For now, change umm_init to default to IRAM.
Define UMM_INIT_USE_ICACHE to move to ICACHE and free up IRAM.

Added some experimental code that may indirectly support PUYA.

Note, until this issue is resolved, that HWDT Stack Dump is not
going to work with PUYA flash.

* typo

* Finalize fix for PUYA flash and preSDK use of Cache_Read_Enable.

This resolves the exception 0 issue with PUYA flash when using
flash/ICACHE for umm_init and/or using HWDT Stack Dump.

Author: mhightower83

cores/esp8266/hwdt_app_entry.cpp

@@ -277,6 +277,7 @@
 #include <esp8266_peri.h>
 #include <uart.h>
 #include <pgmspace.h>
+#include "umm_malloc/umm_malloc.h"
 #include "mmu_iram.h"
 
 extern "C" {

cores/esp8266/mmu_iram.cpp

@@ -197,9 +197,38 @@ extern void Cache_Read_Enable(uint8_t map, uint8_t p, uint8_t v);
 #endif  // #if (MMU_ICACHE_SIZE == 0x4000)
 
 /*
- * This wrapper is for running code from IROM (flash) before the SDK starts.
+ * This wrapper is for running code early from IROM (flash) before the SDK
+ * starts. Since the NONOS SDK will do a full and proper flash device init for
+ * speed and mode, we only do a minimum to make ICACHE functional, keeping IRAM
+ * use to a minimum. After the SDK has started, this function is not needed and
+ * must not be called.
  */
 void IRAM_ATTR mmu_wrap_irom_fn(void (*fn)(void)) {
+  // Cache Read must be disabled. This is always the case on entry when called
+  // from the right context.
+  // Cache_Read_Disable();
+
+  // The SPI_CS_SETUP parameter has been observed set by RTOS SDK and NONOS SDK
+  // as part of flash init/configuration. It may be necessary for some flash
+  // chips to perform correctly with ICACHE hardware access. Turning on and
+  // leaving it on should be okay.
+  //
+  // One SPI bus clock cycle time is inserted between #CS active and 1st SPI bus
+  // clock cycle. The number of clock cycles is in SPI_CNTRL2 SPI_SETUP_TIME,
+  // defaults to 1.
+  SPI0U |= SPIUCSSETUP; // SPI_CS_SETUP or BIT5
+
+  // phy_get_bb_evm is the key function, called from fix_cache_bug in the NONOS
+  // SDK. This addition resolves the PUYA Flash issue with exception 0, when
+  // early Cache_Read_Enable is used.
+  extern uint32_t phy_get_bb_evm(void); // undocumented
+  phy_get_bb_evm();
+
+  // For early Cache_Read_Enable, only do ICACHE_SIZE_16. With this option,
+  // Cache_Read_Disable will fully restore the original register states. With
+  // ICACHE_SIZE_32, one bit is missed when disabling. Leave the full access
+  // calls for the NONOS SDK.
+  // This only works with image slice 0, which is all we do presently.
   Cache_Read_Enable(0, 0, ICACHE_SIZE_16);
   fn();
   Cache_Read_Disable();

cores/esp8266/umm_malloc/umm_malloc_cfgport.h

@@ -27,7 +27,7 @@
  */
 
 /*
- * -DUMM_INIT_USE_IRAM
+ * -DUMM_INIT_USE_ICACHE
  *
  * Historically, the umm_init() call path has been in IRAM. The umm_init() call
  * path is now in ICACHE (flash). Use the build option UMM_INIT_USE_IRAM to
@@ -37,10 +37,16 @@
  * app_entry_redefinable() in core_esp8266_app_entry_noextra4k.cpp for an
  * example of how to toggle between ICACHE and IRAM in your build.
  *
- * The default is to use ICACHE.
+ * ~The default is to use ICACHE.~
+ * For now revert default back to IRAM
+ * define UMM_INIT_USE_ICACHE to use ICACHE/IROM
  */
-// #define UMM_INIT_USE_IRAM 1
-
+#ifdef UMM_INIT_USE_ICACHE
+#undef UMM_INIT_USE_IRAM
+#else
+#undef UMM_INIT_USE_IRAM
+#define UMM_INIT_USE_IRAM 1
+#endif
 
 /*
  * Start addresses and the size of the heap